Guys, I just quantised qwen3 tts

[FNsi]

https://huggingface.co/Jahaz/Qwen3-tts-0.6b-gguf-for-koboldcpp

Having fun!

[FNsi]

Uploaded a smaller tokenizer with some iq3s tensors, and it works well (203M)

[LostRuins]

nice, i didn't try but more quants is always good to have

[FNsi]

nice, i didn't try but more quants is always good to have

😂😂haha. I'd think to make smaller tts model, but cannot be done with the same steps, made an iq3s but stuck in inference, I am gonna test it again later...